arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2102.11529cs.AI
收藏

可微逻辑机

Differentiable Logic Machines

Huawei Noah’s Ark Lab(华为诺亚方舟实验室);Shanghai Jiao Tong University(上海交通大学);IT University of Copenhagen(哥本哈根信息技术大学);Tianjin University(天津大学)

Matthieu Zimmer, Xuening Feng, Claire Glanois, Zhaohui Jiang, Jianyi Zhang, Paul Weng, Dong Li, Jianye Hao, Wulong Liu

AI总结:

本文提出可微逻辑机(DLM),通过为谓词分配权重的连续松弛架构,结合梯度下降训练、演员-评论家与增量训练,同时解决可解释的归纳逻辑程序设计和强化学习问题,并在成功率、奖励和可扩展性上优于现有方法。

摘要:

推理、学习与决策的整合是构建更通用人工智能系统的关键。作为朝此方向迈出的一步,我们提出了一种新颖的神经-逻辑架构,称为可微逻辑机(DLM),它能够同时解决归纳逻辑程序设计(ILP)和强化学习(RL)问题,其中解可以被解释为一阶逻辑程序。我们的方案包含若干创新。首先,与大多数先前的神经-逻辑方法不同,我们的架构通过为谓词而非规则分配权重,定义了一阶逻辑程序空间的一个受限但具有表达力的连续松弛。其次,借助这种可微架构,我们提出了若干基于梯度下降的(监督式和强化学习)训练过程,能够恢复出完全可解释的解(即逻辑公式)。第三,为加速强化学习训练,我们还设计了一种新颖的评论家架构,使演员-评论家算法得以实现。第四,为解决困难问题,我们提出了一种增量训练过程,能够逐步学习逻辑程序。与最先进的(SOTA)可微ILP方法相比,DLM成功解决了所有考虑的ILP问题,并且成功种子的比例更高(最高达3.5倍)。在强化学习问题上,在不要求可解释解的情况下,DLM在奖励方面优于其他不可解释的神经-逻辑强化学习方法(最高达3.9%)。在强制要求可解释性时,DLM能够解决更困难的强化学习问题(例如排序、路径)。此外,我们展示了深度逻辑程序可以通过增量监督训练来学习。除了这一优异性能外,DLM在内存和计算时间方面也具有良好的可扩展性,尤其是在测试阶段,它能够处理比SOTA多得多的常量(超过2倍)。

英文摘要:

The integration of reasoning, learning, and decision-making is key to build more general artificial intelligence systems. As a step in this direction, we propose a novel neural-logic architecture, called differentiable logic machine (DLM), that can solve both inductive logic programming (ILP) and reinforcement learning (RL) problems, where the solution can be interpreted as a first-order logic program. Our proposition includes several innovations. Firstly, our architecture defines a restricted but expressive continuous relaxation of the space of first-order logic programs by assigning weights to predicates instead of rules, in contrast to most previous neural-logic approaches. Secondly, with this differentiable architecture, we propose several (supervised and RL) training procedures, based on gradient descent, which can recover a fully-interpretable solution (i.e., logic formula). Thirdly, to accelerate RL training, we also design a novel critic architecture that enables actor-critic algorithms. Fourthly, to solve hard problems, we propose an incremental training procedure that can learn a logic program progressively. Compared to state-of-the-art (SOTA) differentiable ILP methods, DLM successfully solves all the considered ILP problems with a higher percentage of successful seeds (up to 3.5$\times$). On RL problems, without requiring an interpretable solution, DLM outperforms other non-interpretable neural-logic RL approaches in terms of rewards (up to 3.9%). When enforcing interpretability, DLM can solve harder RL problems (e.g., Sorting, Path) Moreover, we show that deep logic programs can be learned via incremental supervised training. In addition to this excellent performance, DLM can scale well in terms of memory and computational time, especially during the testing phase where it can deal with much more constants ($>$2$\times$) than SOTA.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑